Portions of this patent application contain materials that are subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office. All other rights are expressly reserved.
This invention relates, in general, to distributed computer networks and more specifically to distributed network directory and naming services.
With the tremendous growth of data processing by means of independent, localized data processing devices, such as personal computers and mini computers, data networks have evolved to connect together physically-separated devices and to permit digital communication among the various devices connected to the network.
There are several types of networks, including local area networks (LANs) and wide area networks (WANs). A LAN is a limited area network and data devices connected to a LAN are generally located within the same building. The LAN typically consists of a transmission medium, such as a coaxial cable or a twisted pair which connects together various computers, servers, printers, modems and other digital devices. Each of the devices, which are collectively referred to as xe2x80x9cnodesxe2x80x9d, is connected to the transmission medium at an address which uniquely identifies the node and is used to route data from one node to another. A node which provides resources and services is called a xe2x80x9cserverxe2x80x9d node and a node which uses the resources and services is called a xe2x80x9cclientxe2x80x9d node. A WAN generally encompasses a much larger area and may involve common carrier connections such as telephone lines.
LANs and WANs are often connected together in various configurations to form xe2x80x9centerprisexe2x80x9d networks which may span different buildings or locations or extend across an entire continent. Enterprise networks are convenient for several reasons: they allow resource sharingxe2x80x94programs, data and equipment are available to all nodes connected to the network without regard to the physical location of the resource and the user. Enterprise networks may also provide reliability by making several redundant sources of data available. For example, important data files can be replicated on several storage devices so that, if one of the files is unavailable, for example, due to equipment failure, the duplicate files are available.
One of the most important characteristics of enterprise networks is that they have the capability of bringing a large and sophisticated set of services to all of the attached users for a reasonable cost. However, for the users to exploit the network potential, they must be able to identify, locate and access the network resources. When a network is small, locating and accessing the available services is relatively simple, but networks are growing larger and there are many networks that presently very large. Thousand node networks are common and million node networks are on the horizon.
An example of a very large network is the INTERNET network, which is used by some of the largest public and private organizations. Much of the power of this type of network goes unused simply because the users are either unaware of the facilities available to them or they find the methods of accessing the facilities difficult or confusing. Consequently, in order to assist users in locating and accessing network resources, many existing networks today utilize network directory or naming services which accept a resource identifier or name from a user and locate the network address that corresponds to the desired network resource.
For example, the entered identifier or name can be xe2x80x9cdescriptivexe2x80x9d and specify a resource by describing enough of its attributes to distinguish it from other resources. Such descriptive names are most useful to human users who are searching the network for a resource that meets certain specified criteria, but they are also require the most computing resources and are often difficult to distribute effectively. There presently exist a number standards for such descriptive name services. For example, the Consultative Committee on International Telephony and Telegraphy (CCITT) and the International Standards Organization (ISO) have developed a standard for a descriptive name service known as X.500
Naming and directory services (these will be referred to together as xe2x80x9cdirectory servicesxe2x80x9d hereafter) are presently implemented in a variety of ways. The simplest implementation is to use a single, centralized database contained in a local server node to hold a list of names and corresponding network addresses. An example of such a localized directory service is shown in FIG. 1. FIG. 1 illustrates a computer network arranged in a xe2x80x9cclient-serverxe2x80x9d configuration comprising a plurality of client nodes 106, 108, 120, 122 and 128 which may, for example, be workstations, personal computers, minicomputers or other computing devices on which run application programs that communicate over various network links including links 102, 110, 116, 126 and 136 with each other and with server nodes, such as nodes 100, 112, 124, 132 and 138. The server nodes may contain specialized hardware devices and software programs that can provide a service or set of services to all or some of the client nodes. The client nodes are the users of the various network services which, in turn, are provided by the server nodes
Typically, the centralized directory service database 104 is located in one of the server nodes, such as node 100. A client node, such as client node 108, can access the directory service by connecting to server node 100, entering a resource identifier or name and retrieving the network address of the associated service. By means of conventional database techniques, a client node may be able to search over the database in order to locate a given resource. In addition, many directory services support browsing by using partial name descriptions, xe2x80x9cwild cardsxe2x80x9d and placeholders.
Such centralized directory services with single databases work well in small networks where the number of network addresses is small. However, in larger networks, it is often not feasible to store all the resource identifiers in one central location. Further, a single database represents a single point of failure which can disable the entire network. In addition, a centralized database often suffers from poor performance. For example, while it may be relatively efficient for a local client, such as client 108, to connect to server 100 and access database 104, a remote client, such as client 120, which must link through several servers, 124 and 112, along with a xe2x80x9cgatewayxe2x80x9d link 116, will incur a significant amount of network overhead and the overall system xe2x80x9ccostxe2x80x9d of the access will be high. With a large number of remote access attempts, directory service provider 100 can quickly become both a processing and communication bottleneck for the entire network.
In order to overcome these problems, additional prior art techniques have been developed which distribute the database data over multiple locations. Such a system is shown schematically in FIG. 2. FIG. 2 depicts a client-server type of network which is similar to that shown in FIG. 1. In particular, elements which correspond in the two figures have corresponding numeral designations. For example, client 108 in FIG. 1 is similar to client 208 in FIG. 2. The difference between the two networks is that the directory service database has been replicated in a number of the server nodes. For example, server node 200 contains a directory service database 204 as do server nodes 212 (database 214), server node 232 (database 230) and server node 224 (database 218). There are a number of prior art methods for replicating the data in each of the databases. Some systems replicate each resource identifier individually in each database, other systems replicate the entire database. Still other systems replicate individual nodes or limit replication by partitioning the database in some manner.
The distributed system shown in FIG. 2 avoids the problems associated with the centralized database. Since the data is replicated, there is no single point of failure and, since the data is usually available on a nearby server node, there are no xe2x80x9cremotexe2x80x9d client nodes and network overhead is greatly reduced.
However, the distributed system has its own problems. For example, some method must be used to insure data consistency if multiple sources can update the databases. Some systems force data consistency by keeping all copies of the data tightly synchronized in a manner similar to a conventional database system. Other system insure data integrity by means of conventional concurrency arbitration schemes.
Such distributed naming and directory services are effective on homogeneous networks in which the same access methods and protocols apply over the entire network. In this case, a consistent set of names and rules can be developed to permit location and access of various resources with relative ease. However, many large networks are heterogeneousxe2x80x94not only do the networks comprise many types of different computers, including work stations, personal computers, mini-computers, super-computers and main frames, but the network itself is often composed of many independent smaller networks which are connected together by interfaces called xe2x80x9cgatewaysxe2x80x9d. These smaller networks may have their own access methods and protocols. Further, the heterogeneous construction and organization of these large networks does not lend itself to central control and management which could dictate common methods and protocols.
In many large networks which are comprised of a set of smaller networks which are connected together, each of the underlying separate networks may have its own different directory service utilizing a specific protocol. In this type of network a user may have to be familiar with each network directory service protocol and may have to shift from protocol to protocol as searches are performed from network to network. Consequently, in such a heterogeneous network, one of the main difficulties in accessing network resources arises from a lack of a consistent globally-accessible directory of network resources which can operate over heterogeneous networks without involving the user in the details and the protocol involved in accessing each of these separate networks.
Accordingly, it is an object of the present invention to provide a communications directory service which provides a single globally accessible directory service which is capable of interacting with various existing directory services and other services with existing and future which are provided on a network.
The foregoing problems are solved and the foregoing objects are achieved in one illustrative embodiment of the invention in which a communications directory service is located in each node of the network. The communications directory service includes a tree structure to which existing directory services and other network services can be added. The tree structure has a plurality of nodes each of which includes specific methods that query and browse the associated directory service if such actions are supported by the underlying service. The communications directory service further includes shared libraries which store a service object associated with each service offered on the network. The service object, in turn, includes the service exchange address and communication link configuration information. A client desiring to access a remote service retrieves the appropriate service object from the communications directory service and uses the service object to set up the communications path.
In one embodiment of the invention, each node uses a reconfigurable protocol stack to establish network connections to remote nodes. The communications directory service stores a set of stack definitions which allow the reconfigurable stack to be set up for a particular communication link. Each service object corresponding to a particular service, contains reference to one or more stack definitions for communication links appropriate to that service. When a client retrieves the service object, one of the stack definitions is selected based on criteria such as quality of service or availability of the link.